Dylan Sam

weakly supervised learning (2) learning theory (1) image classification (1) adversarial learning (1) semi-supervised learning (1) algorithmic fairness (1) feature selection (1) weak supervision (1) empirical risk minimization (1) majority vote (1) label noise (1) multi-class classification (1) rademacher complexity (1) label aggregation (1) variational approximation (1) risk estimation (1) observational study (1) loss function (1) noisy label (1) unobserved confounding (1)